`stat_multcomp()` would it be possible to support additional groupings?

[aphalo]

Say factorial experiments? This most likely could be handled by precomputing and using 'ggpp'. In such cases examples in the blog gallery could be written. Related to #30 and could be solved using a shared approach.